Act address landing: DDD-address-01..05, DDD-dec-38, pin advanced to v5.15.1

core/claims/DDD-address-01.yaml

@@ -0,0 +1,52 @@
+format: 1
+id: DDD-address-01
+kind: conceptual
+statement: |
+  Act occurrences carry a time index beside their address, given by causal dependency — occurrence B depends on occurrence A iff a state change produced by A is in B's ground reading — so that the index is the resulting order over state changes rather than wall-clock, dependent occurrences at one address are distinct in time, and independent ones may be concurrent.
+status: projected
+region: |
+  Diachronic — orders occurrences across acts, so it files with the projection (DDD-dec-09's
+  charter). Presupposes upstream DDD-ground-06 (pinned at the release that carries it). The
+  index is kept out of the address because an address containing time would make every
+  occurrence unique and leave nothing to recur at, dissolving behaviour (pre-registration
+  criterion K5). Dependency is read off ground readings, so it inherits their declared-space
+  limit (DDD-ground-05, DDD-ground-02): a state change that reaches B through ground nobody
+  declared does not order B after A.
+evidence:
+  - kind: observation
+    ref: "meta/sessions/2026-09-27-act-address/results-act-address-dry-run.md — T2a"
+    note: |
+      Dry run: retry attempts 1 to 3 of one call share an address and are ordered by dependency
+      alone — each depends on the previous one failing. Read under the amended definition, the
+      previous attempt's failure is in each attempt's ground reading; the dry run predates the
+      amendment and did not record readings. Written inside the framework; instrument check, not
+      evidence for the claim.
+falsifier: |
+  Two dependent occurrences whose order cannot be recovered from the state changes in their ground readings, or a behaviour-change detection that needs a wall-clock distinction the causal order cannot supply.
+test: |
+  T2b (first Phoenix slice): at least two dependent occurrences share an address and are separated only by the time index, per the fixed pre-registration.
+breaks: |
+  Invalidation fan-out as a forward query (the acts after a superseded decision in causal order), and as-of decay's anchor at the act.
+owner: none
+changed: 2026-09-28, act address landing session (GATE 3, on the GATE 1 rulings)
+notes: |
+  Emil originated the extension ("address plus distance in time makes a real map"); ruled R-4,
+  2026-09-27. The causal form, and keeping the index outside the address, were Claude's
+  proposals, accepted in the ruling.
+
+  Amended at the landing session's GATE 1 (Emil, ruling 1, 2026-09-28). The draft left
+  "dependency" undefined. It is now defined over ground readings: B depends on A iff a state
+  change produced by A is in B's ground reading, and the time index is the resulting order over
+  state changes. This is the same line upstream DDD-ground-06's repaired falsifier draws, so the
+  address, the ground readings and the time index are read from one record.
+  UNVERIFIED — Emil review: the reading/verdict-input line — that a dependency always passes
+  through a ground reading and never through a verdict alone. Verdict-triggered decisions
+  (escalation rules) are the pressure case, carried as a T2b test item: they anchor at the
+  following act, whose reading contains the verdict (GATE 2 ruling).
+
+  Rule 1 candidate, adjudicated: kept joined (GATE 1 ruling 4). The definition of dependency and
+  the index it induces are one object; the "so that" clauses are consequences.
+
+  UNVERIFIED — Emil review: that this does not conflict with ground-axes holding note rev 18's
+  argument that time is weak as a coordinate. That argument concerns the age of ground
+  readings; this claim concerns an act's position in dependency order.

core/claims/DDD-address-02.yaml

@@ -0,0 +1,23 @@
+format: 1
+id: DDD-address-02
+kind: normative
+statement: |
+  Every act occurrence is its own act, and reuse of code across occurrences is a derivation-layer relation — many occurrences pulling one component — never an identity between acts.
+status: projected
+region: |
+  Diachronic — relates occurrences across acts. Governs the act map (a tree of occurrences) and
+  places reuse in the derivation record, consistent with forward-only derivation: acts pull
+  components, and components never define acts.
+evidence: []
+falsifier: |
+  A governance question answerable only by treating a reused component as a single act — a case where the occurrence map, with reuse in the derivation layer, cannot state what the identity reading states.
+test: |
+  Counterexample search in T2b: every reused component in the first Phoenix slice is represented as derivation edges from distinct occurrences without loss of any governing decision.
+breaks: |
+  DDD-address-03 and the forward fan-out query; the method's refusal to read acts back from code.
+owner: none
+changed: 2026-09-28, act address landing session (GATE 3, on the GATE 1 rulings)
+notes: |
+  Emil's ruling R-6, 2026-09-27, in his words: "our choice to reuse the same act in the code is an
+  optimization not the truth." Filed as a normative claim: it is a governing choice about
+  representation, and its test is utility.

core/claims/DDD-address-03.yaml

@@ -0,0 +1,29 @@
+format: 1
+id: DDD-address-03
+kind: empirical
+statement: |
+  A change to a component pulled by several act occurrences is an escaped decision at every pulling occurrence that the changing decision's applicability predicate does not cover.
+status: projected
+region: |
+  Diachronic, over the derivation record. Where occurrences replicate a decision instead of
+  sharing a component, the fan-out is the decision's predicate match set, not a component's pullers
+  (dry-run finding F-3, accepted as [PROPOSED] at the landing session's GATE 1). Only the
+  component case is claimed here.
+evidence: []
+falsifier: |
+  Derivation records in which off-predicate pulling occurrences of changed components show no more behaviour change or incident than occurrences that did not pull the component.
+test: |
+  Phoenix: compute each component change's off-predicate fan-out forward from the derivation record, and compare behaviour change at those occurrences against a matched non-pulling set.
+breaks: |
+  The rationale for showing a component's fan-out to the person changing it before the change, and the pricing of reuse as seam risk.
+owner: none
+changed: 2026-09-28, act address landing session (GATE 3, on the GATE 1 rulings)
+notes: |
+  Emil's observation, 2026-09-27: one act often demands a change to a single function, incurring
+  sudden changes in hundreds of others. The seam reading is Claude's, flagged: a component shared
+  by k occurrences is a seam spanning k, decaying at the conjunction of k persistence
+  probabilities. UNVERIFIED — the seam-decay result has no canonical home: the landing session's
+  GATE 1 survey found no claim, term, core document or meta note stating it in either repository
+  at 361e88f / cba63ef. Filed with the reliance carried as a successor item (GATE 1 ruling 6).
+  Only the notes and the "pricing of reuse as seam risk" limb of breaks lean on it; the statement
+  does not. No other claim is cited as support for it.

core/claims/DDD-address-04.yaml

@@ -0,0 +1,23 @@
+format: 1
+id: DDD-address-04
+kind: conceptual
+statement: |
+  Two act occurrences coincide when their ground positions are equal, whatever their layer paths.
+status: projected
+region: |
+  Diachronic — compares occurrences. Needed because the layer path is part of the address
+  (DDD-ground-06), so occurrences under different parents never share a full address.
+evidence: []
+falsifier: |
+  Two occurrences at one ground position under different parents whose comparison misleads — behaviour that differs only because the parents impose different tolerance, read as a ground shift.
+test: |
+  T2b: every cross-parent coincidence in the first Phoenix slice is checked for inherited-tolerance differences (dry-run finding F-2).
+breaks: |
+  DDD-address-05's cross-parent grain.
+owner: none
+changed: 2026-09-28, act address landing session (GATE 3, on the GATE 1 rulings)
+notes: |
+  Ruled R-7, 2026-09-27, on Claude's proposal. The falsifier names the known risk: dry-run finding
+  F-2, that the layer carries inherited tolerance, so cross-parent coincidence may compare unlike
+  acts. F-2 is held, not accepted (landing session GATE 1 ruling 7), and its latency figure is
+  struck; the test checks for the risk rather than presupposing it.

core/claims/DDD-address-05.yaml

@@ -0,0 +1,25 @@
+format: 1
+id: DDD-address-05
+kind: conceptual
+statement: |
+  Behaviour is read at two grains: at full address within one parent, and at ground position across parents.
+status: projected
+region: |
+  Diachronic. Depends on a registered behaviour term, which the Gate 0 bundle queues; until that
+  term lands, "behaviour" here carries the working definition (recurring acts at one address
+  under the same relative ground) and nothing more.
+evidence: []
+falsifier: |
+  A shared-component change that alters behaviour across parents but is invisible at both grains, or a grain that adds no detection over the other on the Phoenix record.
+test: |
+  Phoenix: component changes are detected at the ground-position grain as simultaneous behaviour change across parents.
+breaks: |
+  Behaviour change as the detection channel for DDD-address-03's escapes.
+owner: none
+changed: 2026-09-28, act address landing session (GATE 3, on the GATE 1 rulings)
+notes: |
+  Ruled R-7, 2026-09-27. Filed with the dependency named (landing session GATE 1 ruling 5): the
+  Gate 0 behaviour bundle had not been filed in either repository at 361e88f / cba63ef, and no
+  behaviour term is registered. "Behaviour" here carries only the working definition in region;
+  "relative ground" is likewise unregistered. When the behaviour term lands, this claim is
+  checked against it (successor item).

core/decisions/DDD-dec-38.yaml

@@ -0,0 +1,69 @@
+format: 1
+id: DDD-dec-38
+kind: decision
+statement: |
+  The projection files the act address ruling's diachronic companions — the causal time index, occurrence uniqueness, shared-component escape, the coincidence grain and the two behaviour grains (DDD-address-01 to 05) — as projected claims resting on upstream DDD-ground-06, and advances the upstream pin to the release that carries it, with DDD-ground-06, term:act-address and DDD-ground-01 pinned.
+principal: Emil
+made: 2026-09-27, act address session; filed 2026-09-28 at the act address landing session (GATE 3, on the GATE 1 and GATE 2 rulings); the advance completed at GATE 4, 2026-09-28, against v5.15.1 on the GATE 3 ruling
+resolution: |
+  Final. The GATE 4 sequence completed as ruled. Upstream PR #24 was accepted (merge 13128a7,
+  tagged v5.15.0), but it merged one commit behind the GATE 2 fix. The fix shipped as v5.15.1
+  (PR #25, merge 1cb3819, ratified head 7266306 merged exactly). The ref was bumped from the
+  staged branch to v5.15.1, not v5.15.0, on Emil's GATE 3 ruling. That change of target is a
+  recorded deviation (meta/sessions/2026-09-27-act-address-landing/gate4-target-deviation.md,
+  committed before the bump). The firing did not deviate: ref first, hashes second, it showed
+  74 pins, 0 E12/E13/W5, exactly one W6 (DDD-frame-08, cf73b307… -> ef5d5faf…, both digests as
+  predicted) and W7 1. After re-instrumentation it showed 74 pins with 0 drift. The downstream
+  PR stands for Emil's acceptance.
+
+  The upstream ruling is DDD-dec-37. This decision records the downstream half. The address
+  itself is synchronic and lives upstream. Everything that compares or orders occurrences
+  persists between acts and lives here, per DDD-dec-09 and meta/repo-topology.md, the same split
+  the seam took.
+
+  ## What files
+
+  - DDD-address-01 — the causal time index, amended at GATE 1 (ruling 1): B depends on A iff a
+    state change produced by A is in B's ground reading; the index is the resulting order over
+    state changes. The reading/verdict-input line is flagged UNVERIFIED — Emil review.
+  - DDD-address-02 — every occurrence is its own act; reuse is a derivation relation.
+  - DDD-address-03 — shared-component escape. The seam-decay reading in its notes has no
+    canonical home and is carried as a successor item (ruling 6).
+  - DDD-address-04 — coincidence on ground position. Its known risk is dry-run finding F-2,
+    which is held (ruling 7).
+  - DDD-address-05 — the two behaviour grains, filed with the unregistered behaviour term named
+    as a dependency (ruling 5).
+
+  DDD-address-01 and DDD-ground-06 were kept joined under rule 1 (ruling 4).
+
+  ## The pin advance
+
+  Staged against the upstream landing branch and bumped to the tag only on Emil's acceptance of
+  the upstream PR, per the DDD-dec-10/16/18/25/28 pattern. The predicted firing was committed
+  before the ref moved (meta/sessions/2026-09-27-act-address-landing/gate3-prediction.md). The
+  advance runs from v5.13.0, so it crosses v5.14.0 and v5.14.1 as well: it carries v5.14.0's
+  published W6 on DDD-frame-08, and nothing else from the existing pins.
+
+  Three pins are added. DDD-ground-06 and term:act-address are the basis of the five claims and
+  do not exist before the release that carries them, so they could not be pinned earlier.
+  DDD-ground-01 is this decision's basis and was not pinned: a basis cited without a pin can move
+  silently (DDD-agent-01). DDD-dec-37 is not pinned, because the checker indexes upstream claims
+  and terms, not decisions.
+
+  ## Rulings on the dry run (GATE 1, ruling 7)
+
+  E-1 is accepted as a correction record. F-3 is accepted as [PROPOSED]. The T2a act map is
+  accepted, with a1's gateway axis recorded as not read. F-1 is held as a T2b test item: an act
+  that never reads an axis a governing predicate uses. F-2 is held and its 14 s figure is struck.
+  The session record lands exactly as bundled; the rulings live in the landing session's
+  gate1-rulings.md, not in the record they rule on.
+basedOn:
+  - DDD-ground-06
+  - DDD-dec-37
+  - DDD-ground-01
+notes: |
+  New claim area `address`. Areas are cheap and renumbering is forbidden (spec/claim-format.md
+  §3). The area name was confirmed free at the landing session's GATE 1.
+
+  Phoenix is the code name of the first independent application case. It is named here and in
+  the claims' tests. Upstream names the case neutrally (GATE 1 ruling 2).

graph/upstream.yaml

@@ -100,7 +100,23 @@ upstream:
   # none removed. The same session's branch-only supersession of DDD-frame-08 (GATE 1 ruling C)
   # was predicted to contribute nothing to this advance — the tag does not carry it — and
   # contributed nothing; its firing belongs to the next advance.
-  ref: v5.13.0                      # tag, never a branch
+  # Advanced v5.13.0 -> v5.15.1 at the act address landing session (2026-09-28), on Emil's
+  # acceptance of upstream PR #24 (merge 13128a7, tagged v5.15.0) and PR #25 (merge 1cb3819,
+  # tagged v5.15.1); the advance is DDD-dec-38 here, DDD-dec-37 upstream. Staged against the
+  # landing branch per the DDD-dec-10/16/18/25/28 pattern, then bumped to the tag. DEVIATION,
+  # recorded before the bump at meta/sessions/2026-09-27-act-address-landing/
+  # gate4-target-deviation.md: the prediction (gate3-prediction.md, committed at ef0d444 before
+  # the ref first moved) targeted v5.15.0; v5.15.0 was cut from e484ca5 without the GATE 2 fix
+  # cc0abf8, which ships as v5.15.1, and Emil ruled the pin names v5.15.1. Notes and i18n only
+  # separate the two, so the restated prediction is identical: E12 0 - E13 0 - W5 0 - W6
+  # exactly 1 (DDD-frame-08, cf73b307... -> ef5d5faf..., v5.14.0's published prediction) - W7 1,
+  # unchanged. Three pins added: DDD-ground-06, term:act-address, DDD-ground-01.
+  # Verified at the bump, ref advanced first and hashes re-instrumented second so the firing was
+  # observed rather than assumed: 74 pins resolved, 0 basis-loss, EXACTLY ONE content-drift —
+  # DDD-frame-08, both digests as written down beforehand — 1 shadowed id. The prediction held
+  # in every limb. After re-instrumenting DDD-frame-08: 74 pins, 0 basis-loss, 0 content-drift,
+  # 1 shadowed id.
+  ref: v5.15.1                      # tag, never a branch
   pins:
     # basis claims for DDD-dec-04 (the split decision itself rests on these)
     - id: DDD-frame-01
@@ -282,7 +298,7 @@ upstream:
     # stated without a citation; the citation is the fix). Same ref (v5.7.0), no advance.
     - id: DDD-frame-08
       status_at_pin: projected
-      content_hash: sha256:cf73b307ede964603c2eb36863f4a4b7874894350e05ae467b300fcf54aa3590
+      content_hash: sha256:ef5d5fafedb588994141c8f4df504c7589ab3ea7b5257f3ea97f37ba3db07ff7
     - id: term:acceptance-predicate
       status_at_pin: settled
       content_hash: sha256:a3aaae2e710ffb955b30e2e4b0f8371b72c1ada7816ac698dab5e2310ee29e78
@@ -357,3 +373,16 @@ upstream:
     - id: DDD-measure-17
       status_at_pin: projected
       content_hash: sha256:164a18072145bc4a0cdb3c6e1ca8890fc64c1b36396c73be5ca8bd45c66114e5
+    # [PROPOSED] Act address landing (2026-09-28, DDD-dec-38), staged with the v5.15.0 advance.
+    # DDD-ground-06 and term:act-address are the basis of DDD-address-01..05; they do not exist
+    # before v5.15.0, so pinning them earlier would fire E12. DDD-ground-01 is DDD-dec-38's
+    # basis and was unpinned (DDD-agent-01). Hashes taken at upstream cc0abf8.
+    - id: DDD-ground-06
+      status_at_pin: projected
+      content_hash: sha256:87a6adab30551a4f464cdc1020ae0155f22d6a86d3baac30917894a1cb3b6609
+    - id: term:act-address
+      status_at_pin: draft
+      content_hash: sha256:d66ddee2ca634dc8ab84d96798b5ca0af36403551c826fb3affd86758f1df00c
+    - id: DDD-ground-01
+      status_at_pin: projected
+      content_hash: sha256:eb4679f9878a220b8a4c39e785b1d3cd9f6ad7eb717a02d1eff6ff97ef359d93

meta/sessions/2026-09-27-act-address-landing/bootstrap.md

@@ -0,0 +1,21 @@
+# Bootstrap — act address landing
+
+Filled in at arrival and committed with `prompt.md` as the session's first commit.
+
+| Parameter | Value |
+|---|---|
+| Invocation | The contents of `prompt.md`, given verbatim with the bundle attached as `act-address-bundle.zip` |
+| Principal | Emil (sole ratifier) |
+| Bundle path | Unpacked in the session scratchpad as `act-address-bundle/`; outside both repositories |
+| Upstream repo | github.com/mindovermachine-dev/actor-indexed-determination |
+| Upstream base | `361e88f` — head of `main` at arrival; tag `v5.14.1` verified to point at it. No drift from the drafts' base |
+| Downstream repo | github.com/mindovermachine-dev/decision-driven-design |
+| Downstream base | `cba63ef` — head of `main` at arrival. No drift from the drafts' base |
+| Branch, both repos | `claude/act-address-landing-bla6io` (the harness-assigned branch; the bundled template named `claude/act-address-landing`) |
+| Gates | 0 arrival · 1 survey · 2 upstream · 3 downstream · 4 close |
+| Pre-registration hash | `ef34e3bc14ab793800edd8c096e82adf0652ef8bc284d54d52efd20f8bf5f5eb` — verified against the bundled `prereg-act-coordinates.md` at arrival |
+
+## Deviations from the bundled template, recorded at arrival
+
+- **Directory name.** `LANDING-DATE` resolved to `2026-09-27`, the arrival date. The landing therefore shares its date with the session it lands (`2026-09-27-act-address/`); the slugs differ.
+- **Branch name.** The session runs on the branch the harness assigned, which carries a suffix. The bundled template's unsuffixed name was not created.